The primary agent ignored instructions to only review final summaries.
@davidmanheim Do you see these issues in codex too?
In many ways, new models are very self-aware and situationally aware - but not in the one way I'd like them to be; they have trouble delegating without micromanaging, even when explicitly told to do so.
Am I missing something? And/or is this a skills issue?
The primary agent ignored instructions to only review final summaries.
@davidmanheim Do you see these issues in codex too?
In many ways, new models are very self-aware and situationally aware - but not in the one way I'd like them to be; they have trouble delegating without micromanaging, even when explicitly told to do so.
Am I missing something? And/or is this a skills issue?